Bulletin of the World Health Organization

Control of hypertension with medication: a comparative analysis of national surveys in 20 countries

Nayu Ikeda a, David Sapienza b, Ramiro Guerrero c, Wichai Aekplakorn d, Mohsen Naghavi e, Ali H Mokdad e, Rafael Lozano f, Christopher JL Murray e & Stephen S Lim e

a. AXA Department of Health and Human Security, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan.
b. Swedish Family Medicine-Cherry Hill, Seattle, United States of America (USA).
c. Centro de Estudios en Protección Social y Economía de Salud, Universidad Icesi, Cali, Colombia.
d. Ramathibodi Hospital, Mahidol University, Bangkok, Thailand.
e. Institute for Health Metrics and Evaluation, University of Washington, Seattle, USA.
f. National Institute of Public Health, Cuernavaca, Mexico.

Correspondence to Nayu Ikeda (e-mail: ikedan@m.u-tokyo.ac.jp).

(Submitted: 18 March 2013 – Revised version received: 15 July 2013 – Accepted: 23 July 2013 – Published online: 30 September 2013.)

Bulletin of the World Health Organization 2014;92:10-19C. doi: http://dx.doi.org/10.2471/BLT.13.121954


High blood pressure, also known as hypertension, is a major contributor to the global disease burden and was responsible for 7% of all disability-adjusted life years in 2010.1 Moreover, the number of people with uncontrolled hypertension has increased to around 1 billion worldwide in the past three decades.2 As a result, the effective control of hypertension has become a priority for global health policy and, with growing interest in the prevention and control of noncommunicable diseases (NCDs),3 it is vital that health-care systems deliver appropriate interventions for tackling high blood pressure.

The formulation of effective policies for decreasing the burden of uncontrolled hypertension depends on knowledge of the current rate of hypertension control at the population level. Several countries have carried out health examination surveys of nationally representative samples to measure blood pressure and to assess awareness, treatment and control of hypertension in the general population.411 In addition, trends in indicators of hypertension management have been reported in Canada,12 England13 and the United States of America14,15 and coverage of effective treatment for hypertension has been studied in the Islamic Republic of Iran11 and Mexico.16,17

Despite the wealth of data available from national health surveys on the management of hypertension in the general population, it is not always possible to compare published results directly because of differences in survey methods and analytical strategies.18,19 Previous studies have reviewed findings from several national surveys on the crude prevalence of hypertension and on awareness, treatment and control of the condition20,21 but the issue of comparability was left unresolved. Although one study analysed microdata from national examination surveys, it covered only countries in Europe and North America.19,22 Another recent study used regression modelling techniques to aggregate data and improve comparability.2 However, awareness, treatment and control of hypertension were not analysed. Thus, comparative information – as is already available for diabetes and hypercholesterolaemia – is needed to benchmark and compare how health systems perform in controlling hypertension.23,24

Our aim was to provide comparable, comprehensive and consistent evidence on the management of hypertension internationally by analysing data from health surveys in 20 countries that included blood pressure measurement.


Two investigators obtained information on national health examination surveys by systematically searching: (i) published research papers using Medline and Google Scholar; and (ii) health survey databases, such as the Global InfoBase of the World Health Organization (WHO)25 and the STEPwise approach to Surveillance (STEPS) database,26 the European Health Interview and Health Examination Surveys database,27 the Demographic and Health Survey database28 and databases on other web sites identified during the search process. The keywords used were blood pressure, high blood pressure, hypertension, prevalence, treatment, control and national health survey as well as country names. Articles and survey reports were reviewed to determine whether they: (i) involved a random sample of adults in a whole country; (ii) included both sexes; (iii) contained data on blood pressure measurements; and (iv) contained data on the diagnosis of hypertension and on the use of antihypertensive medications.

We identified nationally representative household surveys that met our four criteria for 73 of the 193 WHO Member States: 12 were low-income countries, 38 were middle-income countries and 20 were high-income countries; income was not classified for 3 countries. For these 73 countries, we tried to obtain anonymized individual-level data. We downloaded publicly available data sets for England, the Russian Federation, the United States and 9 countries for which Demographic and Health Survey data were available. For the other 61 countries, we requested data from the institutions that conducted the surveys and we obtained permission to use data from 8 countries. Overall, we obtained individual-level data for 20 countries in which the latest available surveys had been conducted between 1992 and 2011 (Table 1, available at: http://www.who.int/bulletin/volumes/92/1/13-121954). All Demographic and Health Surveys, except those for Bangladesh and South Africa, included only women of reproductive age and their husbands. Data from several different years were available for England, Japan and the United States.

Table 1 summarizes the procedures used to measure blood pressure in the surveys included in this study. In most surveys, blood pressure was measured more than twice but only single measurements were recorded in the surveys carried out in Colombia and in Japan in 1980 and between 1986 and 1999. Digital blood pressure metres were used in England and all countries participating in Demographic and Health Surveys except Uzbekistan. All surveys recorded blood pressure in the right arm of seated participants, except those in South Africa and Turkey.

Individuals were regarded as having hypertension if their systolic blood pressure was equal to or greater than 140 mmHg or they reported current use of blood-pressure-lowering medications. We did not include diastolic blood pressure in our definition of hypertension because prospective studies suggest that systolic pressure is a better predictor of cardiovascular disease risk, particularly in older individuals.46,47 Moreover, systolic pressure rises consistently with age, whereas diastolic pressure increases until the age of 50 to 60 years and starts decreasing thereafter.48 For each individual, we calculated systolic blood pressure as the average of all measurements taken, excluding the first measurement, when three or more measurements were available; when only one or two measurements were available, we used the single or second measurement, respectively.

Among people with hypertension, we estimated the fraction that had been diagnosed with the condition, the fraction treated and the fraction whose condition was controlled using antihypertensive medications. We regarded participants as having been diagnosed with hypertension if they answered “yes” when asked whether a doctor or any other health professional had told them they had high blood pressure. The questions used to determine whether a participant had been diagnosed with hypertension were slightly different in some countries: in all surveys in England, except those performed in 2007 and 2008, two questions were asked to determine whether a diagnosis had been made in a clinical setting; in the German survey, “health professionals” were not specifically mentioned; and, in surveys in the Islamic Republic of Iran, Jordan and Thailand, the recall period was limited to the previous 12 months (Table 2, available at: http://www.who.int/bulletin/volumes/92/1/13-121954). In addition, the 2007 and 2008 surveys in England and the surveys in Japan in all years except 1980, 1990 and 2000 did not ask about a history of hypertension.

We considered hypertensive patients to be receiving treatment if they reported currently using an antihypertensive medication and we regarded a systolic blood pressure less than 140 mmHg as indicating that the hypertension was controlled. Most surveys employed a single yes–no question to ask respondents about their current use of antihypertensive drugs (Table 2). The German survey and Japanese surveys in 1980 and 1990 to 2002 asked about the frequency of medication use and we considered only those who answered “daily” to be receiving treatment. Most surveys did not ask about treatment if the respondent answered “no” to a question about being diagnosed with hypertension. In Colombia, only individuals who had been diagnosed with hypertension two or more times were asked about the use of antihypertensive medications. The South African survey asked for the names of the medications to verify that they really were for lowering blood pressure.

We included all individuals aged 35 to 84 years in the analysis to maximize the overlap of age ranges across surveys. We excluded women who were pregnant or breastfeeding and individuals for whom data on any of the following were missing: systolic blood pressure; a history of hypertension (except for some of the English and Japanese surveys, which did not ask about this item); and the use of antihypertensive medications. We could not identify pregnant women in the Jordanian survey because of a lack of data.

Overall, we compiled a data set covering 173 920 individuals from 20 countries for the cross-sectional analysis and, for the trend analysis, we had data on 63 903 individuals from England, 155 212 from Japan and 30 410 from the United States.

We estimated the age-standardized prevalence of hypertension in individuals aged 35 to 49 years in the latest surveys from 20 countries and, in individuals aged 35 to 84 years, in surveys from 11 countries. In these groups, we also estimated the age-standardized proportion of individuals with hypertension whose condition was diagnosed, treated or controlled by means of antihypertensive drugs. In addition, we examined secular trends in these indicators in the population aged 35 to 84 years in England, Japan and the United States. We derived a reference population for the age standardization of prevalence data by calculating an average for the proportion in each 5-year age group in all countries included in the analysis using population estimates for the year 2000.49 For the age standardization of the proportion of individuals whose hypertension was diagnosed, treated or controlled, we derived a standard hypertensive population from the reference population by using an average estimate of the prevalence of hypertension in each 5-year age group across all countries. We used Stata version 12 (StataCorp. LP, College Station, USA) for the analysis and to adjust for complex survey designs that included stratification, clustering and sample weights.


The age-standardized prevalence of hypertension in individuals aged 35 to 49 years ranged from around 12% in Bangladesh, Egypt and Thailand to around 30% in Armenia, Lesotho and Ukraine (Table 3). In those aged 35 to 84 years, it varied from nearly 20% in Bangladesh to more than 40% in Germany, the Russian Federation and Turkey (Table 3).

In individuals with hypertension aged 35 to 49 years, the age-standardized percentage diagnosed with the condition was highest in the United States, at 83.9%, followed by Uzbekistan and Lesotho. It was under 34% in Albania, Armenia, Colombia, the Islamic Republic of Iran and Turkey (Table 4). The age-standardized percentage whose blood pressure was controlled with medications was 55.7% in the United States, whereas it was less than 10% in Albania, Armenia, Germany, the Islamic Republic of Iran and Turkey (Table 4).

In individuals with hypertension aged 35 to 84 years, the age-standardized percentage whose high blood pressure was diagnosed, treated or controlled was highest in the United States and lowest in Germany, Thailand and Turkey (Table 4). The percentage diagnosed ranged from less than 50% in Thailand and Turkey to 85.3% in the United States (Table 4). The percentage of hypertensive individuals who were diagnosed but not on medication was highest in Turkey, at 20.7%. The percentage whose hypertension was controlled with medications varied from less than 8% in Germany and Turkey to 59.1% in the United States (Table 4).

Our analysis of the secular trend in the age-standardized prevalence of hypertension in individuals aged 35 to 84 years showed that the prevalence was substantially lower in the United States in the early 1990s than it was in England or Japan, the other two countries in which the trend was analysed (Table 5). However, the prevalence increased in the United States during the 1990s, whereas it decreased in England and Japan. Subsequently, the prevalence became comparable across the three countries in the 2000s. The age-standardized percentage of hypertensive individuals whose hypertension was diagnosed, treated or controlled increased over time in all three countries (Table 6, available at: http://www.who.int/bulletin/volumes/92/1/13-121954). The increase was particularly remarkable in England: although the percentage whose hypertension was treated or controlled in England was lower than in Japan in the early 1990s, by the late 2000s, the percentage in England exceeded that in Japan.


To the best of our knowledge, this is the first study to compare the management of high blood pressure in countries from different regions of the world by using individual-level data from nationally representative health examination surveys. Our results suggest that hypertension is not adequately controlled with medication in many parts of the world. This finding has important implications for medical and public health authorities worldwide and should be considered when formulating and implementing programmes for controlling hypertension.

We found that the prevalence of hypertension was substantial in some low- and middle-income countries, which indicates that the condition is an important contributor to the growing burden of NCDs in developing countries. Moreover, blood pressure control in hypertensive individuals was particularly poor in Albania, Armenia, the Islamic Republic of Iran and Turkey, as well as in Germany in the late 1990s. Among former socialist states in Europe and central Asia, the proportion of adults aged 35 to 49 years with hypertension whose condition was controlled was much greater in Azerbaijan, Ukraine and Uzbekistan than in Albania and Armenia. Although few studies of changes in the management of hypertension have been carried out in transitional economies, one investigation showed that, in Armenia, inadequate health care utilization was partly due to the high cost for users.50 A study in Mexico revealed that health system reform during the early 2000s reduced blood pressure in the population by only around one fifth of the maximum potential decrease that could have been achieved with antihypertensive drugs.16,17 Our results confirmed that there is still considerable room for improvement in Mexico. In Thailand, the prevalence of hypertension was lower than in other countries in our study but hypertension control was relatively poor. This finding supports a previous study’s proposal that the screening, treatment and control of hypertension in Thailand should be strengthened.8

The fact that the age-adjusted prevalence of hypertension remained between 35 and 40% throughout the 2000s in England, Japan and the United States suggests that there has been little progress in the primary prevention of the condition in these countries. Moreover, previous studies showed that the prevalence of hypertension increased in the United States in the 1990s,14,15 perhaps partly on account of the increasing number of overweight and obese individuals.14 In England, the substantial decrease in the prevalence of hypertension observed in the 1990s and early 2000s13 seemed to have ceased by the late 2000s. It is a challenge for these highly industrialized countries to resume progress in reducing the prevalence of hypertension, particularly in older individuals.

Of all the countries included in our analysis, the United States showed the highest rates of diagnosis, treatment and control of hypertension. Since the late 1990s, England has been rapidly catching up with the United States. The improvement may partly be explained by the influence of a series of guidelines on the management of hypertension disseminated by the British Hypertension Society and by a pay-for-performance system introduced in the new General Medical Services contract that gives general practitioners an incentive to lower patients’ blood pressure.13 In contrast, progress in controlling hypertension with medications has been relatively slow in Japan. Although substantial efforts have been made in the country to improve the management of hypertension, many physicians may be cautious about aggressively treating the condition, particularly in elderly patients. A fundamental change in physicians’ attitudes may be needed to achieve the same level of hypertension control as has been achieved in England and the United States.

Our study had several limitations. First, although we used the latest available surveys, those of Germany, the Russian Federation and South Africa all dated from the 1990s and were too old for us to draw conclusions about the current state of hypertension management in those countries. Second, although systolic blood pressure is a better predictor of cardiovascular risk, the exclusion of diastolic blood pressure from our definition of hypertension may have led us to underestimate the prevalence of the condition in individuals younger than 50 years. Third, information about diagnosis and treatment was obtained from survey respondents’ answers to questions and may have been affected by recall bias and variations in the wording of questions. Fourth, most surveys employed a question with a yes–no answer to record the use of antihypertensive medications and did not ask about frequency of use. Consequently, our estimates of the proportion of hypertensive individuals being treated may have included those taking medications less than daily and this may have resulted in underestimates of the proportion whose hypertension was under control. Fifth, in this study, we focused on treatment with antihypertensive medications. The prevalence of hypertension and the proportions of hypertensive individuals whose high blood pressure is treated or controlled would be different in countries in which nonpharmacological interventions, such as dietary modification and physical activity, are common. Sixth, although it was a nationally representative sample, the survey sample in Jordan was small. Consequently, the figures for the percentage of hypertensive individuals, particularly older individuals, whose condition was diagnosed, treated or controlled by medication should be interpreted with caution.

The strength of our study is that the use of individual-level data enabled us to apply consistent definitions and analytical methods when comparing estimates of the prevalence, awareness, treatment and control of hypertension across countries and over time. The use of consistent analytical methods is crucial for assessing how well individual health systems are responding to the health needs of the population. Therefore, we believe our study is important as a first attempt at providing consistent and comparable information on the management of hypertension globally.

Considerably more national health examination surveys on high blood pressure have been conducted around the world than on high blood glucose or cholesterol concentrations.23,24 Nevertheless, largely owing to financial constraints, most countries have not carried out surveys that involve blood pressure measurement and very few have established continuous surveys. In our study, we were able to obtain access to individual-level data for only 20 of 73 countries identified. Consequently, our final survey selection was somewhat opportunistic, which may have limited the representativeness of our data and its generalizability to other countries. In addition, opportunities to assess, benchmark and compare the performance of health policies and programmes internationally may have been missed. As has been proposed for the assessment of hypercholesterolaemia,24 there is a clear need for a standardized protocol for implementing blood pressure surveys. Those we identified adopted different ways of measuring blood pressure and many asked respondents themselves to report their history of hypertension and current treatment status. Errors that result from the heterogeneity in survey methods are difficult to correct. In resource-poor settings, the user-friendly techniques employed by Demographic and Health Surveys would be helpful for gathering biomarker data. Use of these techniques should be promoted by the global campaigns on the prevention and control of NCDs planned for coming years.

In conclusion, globally the proportion of hypertensive patients whose disease is treated effectively with medications remains low, especially in some low- and middle-income countries. Since hypertension is the leading risk factor associated with morbidity and mortality from NCDs worldwide, greater efforts should be devoted to improving hypertension control. Countries and different clinical disciplines should work together to adopt a comprehensive approach to the prevention and control of high blood pressure. Effective and affordable drugs are available. It is time to act, since the alternative is expensive and a large portion of the population is affected.


We thank Catherine Claiborne, Majid Ezzati and Emmanuela Gakidou. Rafael Lozano is also affiliated with the Institute for Health Metrics and Evaluation, University of Washington, Seattle, USA.


This study was supported in part by the Bill & Melinda Gates Foundation and by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (grant numbers: 24590785 and 25253051). Nayu Ikeda was the beneficiary of financial support from the AXA Research Fund.

Competing interests:

None declared.